This was interesting but just wanted to note they defined "known" based off of a test where you are asked yes/no on a series of words based on if they are an actual English word or made up. Not remotely if you can accurately define/use them and they didn't go into details on how often people were voting "yes" on words that turned out to not even exist. I would be absolutely shocked in a random sample of American men if even 20% could accurately define "aileron" or "azimuth".
I was also curious about the population of participants. The paper has this to say:
> The test was made available on a dedicated website (http://vocabulary.ugent.be/). Access to the test was unlimited.
> we analyzed the data of 221,268 individuals who completed 265,346 sessions.
That's a pretty large sample, although I wonder how they recruited the participants. It may be in the paper but I couldn't find mention of that in a quick skim.
